acm-knowledge-learner

Updating an ACM knowledge base with live cluster state involves comparing the active cluster against a curated baseline to discover unknown operators, new failure patterns, dependency chains, and certificate issues. This process can be fully automated.

Component discovery for unknown operators works by comparing your live cluster state to the existing knowledge base to identify gaps. It optionally uses MCPs for dependency discovery and component understanding to map new failure patterns.

Yes, you need the oc CLI logged into an ACM hub to discover dependency chains and certificate issues. Optional MCPs can be integrated to further enhance dependency discovery and component understanding.

The best way to automate learning from a cluster and refreshing baselines is to run an automated comparison between the live cluster state and your curated knowledge base to extract new failure patterns and dependency chains.

Yes, you can use neo4j-rhacm alongside acm-source and acm-search as dependencies to map unknown operators and dependency chains, providing deeper component understanding when refreshing the ACM knowledge base.

New failure patterns might not appear in your ACM knowledge base if the live cluster state has not been recently compared against the baseline, or if the oc CLI is not properly logged into the ACM hub to extract current component data.